Semantic Enabled Metadata Framework for Data Grids

We designed a semantic enabled metadata framework using ontology for multi-disciplinary and multi-institutional large scale scientific data sets in a Data Grid setting. There are two main issues we intend to address: data integration for semantically and physically heterogeneous distributed knowledge stores, and semantic reasoning for data verification and inference in such a setting. This framework enables data interoperability between otherwise semantically incompatible data sources, cross-domain query capabilities and multi-source knowledge extraction. In this paper, we present the basic system architecture for this framework, as well as an initial implementation.